|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|
Re: [PLUG] Transfer REALLY large file(s) between Linux and OS X
|
OK, here you go:
Simple rcp from Sun-450 to Dual Xeon /dev/null:
822,190,683 ba0001.tape
time rcp rain-ge0:/xy/F/broom/m6841/data/ba0001.tape /dev/null
3 times to make sure caching or IO does not come into play
time rcp rain-ge0:/xy/F/broom/m6841/data/ba0001.tape /dev/null
real 0m28.521s
user 0m0.090s
sys 0m1.950s
real 0m27.620s
user 0m0.070s
sys 0m1.830s
real 0m28.402s
user 0m0.110s
sys 0m1.430s
Let's say 28 MB per second.
The Sun overhead was about 10% of CPU, which really means about 2/5 of a
CPU since we have 4 in the box. They are SPARC 450Mhz.
Local was about 10% of a single Xeon.
time scp rain-ge0:/xy/F/broom/m6841/data/ba0001.tape /dev/null
ba0001.tape 100% |*********************************| 784 MB 02:30
real 2m35.366s
user 0m0.150s
sys 0m0.820s
ba0001.tape 100% |*********************************| 784 MB 02:31
real 2m35.452s
user 0m0.120s
sys 0m1.070s
ba0001.tape 100% |**********************************| 784 MB 02:33
real 2m38.026s
user 0m0.170s
sys 0m0.700s
The CPU on the Sun is now pinning at 25%. The means 100% of a single
CPU is being used. It can't go any faster. It is moving the
data at a trivial 5.2 MB per second.
I used the report copy time (ie: 2:30) to make sure I did not include any
setup overhead.
On Sunday 19 October 2003 07:11 pm, Tobias DiPasquale wrote:
> One would also then necessarily assume that some benchmarking was in
> order to determine whether or not that overhead is negligable or not for
> differing sizes of transferred data. That's what I was asking for, in
> case that escaped your attention.
___________________________________________________________________________
Philadelphia Linux Users Group -- http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug
|
|